raw byte strings are described in a confusing way

[warsaw]

BPO	13744
Nosy	@gvanrossum, @warsaw, @birkenfeld, @terryjreedy, @pitrou, @merwok
Files	stringByteLiteralBR.docx: Correct update to documentation

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2013-10-13.08:01:56.872>
created_at = <Date 2012-01-09.13:34:21.499>
labels = ['type-feature', 'docs']
title = 'raw byte strings are described in a confusing way'
updated_at = <Date 2013-10-13.08:01:56.870>
user = 'https://github.com/warsaw'

bugs.python.org fields:

activity = <Date 2013-10-13.08:01:56.870>
actor = 'georg.brandl'
assignee = 'docs@python'
closed = True
closed_date = <Date 2013-10-13.08:01:56.872>
closer = 'georg.brandl'
components = ['Documentation']
creation = <Date 2012-01-09.13:34:21.499>
creator = 'barry'
dependencies = []
files = ['25065']
hgrepos = []
issue_num = 13744
keywords = []
message_count = 10.0
messages = ['150936', '150940', '150941', '150942', '150944', '157046', '157047', '157122', '157123', '199667']
nosy_count = 8.0
nosy_names = ['gvanrossum', 'barry', 'georg.brandl', 'terry.reedy', 'pitrou', 'eric.araujo', 'docs@python', 'j.chadwick']
pr_nums = []
priority = 'normal'
resolution = 'out of date'
stage = None
status = 'closed'
superseder = None
type = 'enhancement'
url = 'https://bugs.python.org/issue13744'
versions = ['Python 2.7', 'Python 3.2', 'Python 3.3']

[warsaw]

The lexical analysis documentation says this:

http://docs.python.org/py3k/reference/lexical_analysis.html?highlight=raw%20bytes

"Bytes literals are always prefixed with 'b' or 'B';..."

"Both string and bytes literals may optionally be prefixed with a letter 'r' or 'R';..."

But that would lead you to believe that to get raw byte strings you should use rb"foo". In fact, that's a SyntaxError in Python 2.6+ and Python 3. What *does* work though is br"foo".

Either Python should accept both spellings (harder) or the documentation should make it clear that the 'b' must preceded the 'r'.

[pitrou]

Either Python should accept both spellings

+1. Been annoyed several times by this.

[warsaw]

On Jan 09, 2012, at 03:15 PM, Antoine Pitrou wrote:

Antoine Pitrou <pitrou@free.fr> added the comment:

> Either Python should accept both spellings

+1. Been annoyed several times by this.

The $64k question: is this a new feature or a bug? :)

[pitrou]

The $64k question: is this a new feature or a bug? :)

Most certainly a feature...

[warsaw]

On Jan 09, 2012, at 03:35 PM, Antoine Pitrou wrote:

Antoine Pitrou <pitrou@free.fr> added the comment:

> The $64k question: is this a new feature or a bug? :)

Most certainly a feature...

In that case, since we can only add the new prefixes to 3.3, I still think we
need to fix the documentation to remove the confusion for stable releases.

[jchadwick]

The attached replaces the text for the documentation in 2.4.1 between the lexical definitions table and the escape sequence table. The only change is the following addition to the paragraph on string and byte literals prefixed by 'r' or 'R':

When a byte literal is prefixed with both 'r' or 'R' and 'b' or 'B', the b must precede the r, as in: 'Br', 'bR', or 'BR' and not 'Rb', 'rB', or 'RB'.

[jchadwick]

I uploaded before making the final save, so the first document is incomplete.

(that's embarrassing)

[merwok]

Thanks for the contribution. Could you post your suggested wording as a plain text file?

[merwok]

Actually it’s fine, we already have the text in your message :)

[birkenfeld]

3.3 is now released with both "br" and "rb" allowed, and in 2.7 the current text looks good enough for me.